204 Scientific Proceedings, Royal Dublin Society. 



of any of the lines or bands of a spectrum may be due solely to a particular 

 compound, and not to either of the elements of wliieh it is composed, or to the 

 accidental presence of some other element. 



The apparatus used consisted of a spectrograph, coil, condenser, pump, 

 dryers, pressure gauge, and quartz vacuum tube. Tlie spectrograph was 

 one of Sir Walter Noel Hartley's design, the property of the Royal Society, 

 and kindly lent me for the purposes of this research. Tlie research was also 

 assisted financially by the Government Grant Committee, to enable me to 

 procure the large number of quartz vacuum tubes necessary. The spectro- 

 graph is a one-prism instrument, fitted with quartz lenses of 15 inches focal 

 length. It was carefully adjusted with cadmium line 17 at the angle of 

 minimum deviation, using a barium platino-cyanide screen. This line was 

 then brought into accurate focus on a plate, and the plate rotated until the 

 best possible general effect was obtained from the red to the ultra violet. 

 Through a slight defect in the plate-carrier the lines of the spectra are not 

 vertical ; and at the time of the experiments it was not thought desirable to 

 alter this, as the focus of the instrument was very good, and gave an excellent 

 spectrum from the red to the extreme ultra violet. To remedy this for 

 purposes of reproduction would have entailed a repetition of the whole of the 

 work, which is quite unnecessary for the accurate measurement and identifi- 

 cation of the lines, the scale at the top of the plate being only intended as a 

 rough indication of the position of the lines to facilitate the comparison of the 

 photographs of spectra veith the table of wave-lengths. 



The coil was an Apps coil, giving a 12-inch spark when worked by five 

 storage cells, and the condenser in the secondary circuit consisted of a sheet of 

 glass, with two sheets of tin foil about 15 inches by 18 inches. 



The pump was of the Geryk type, supplied by the Pulsometer Company. 

 To it were fixed drying-tubes of phosphorous peiitoxide and some tubes of 

 caustic potash to absorb any acid vapours that might be liberated from 

 some of the chlorides ; also a tube dipping into mercury, set by the side of a 

 barometric tube, to give a rough indication of the degree of exhaustion. Two 

 or three strokes of the pump easily exhaust vacuum tubes to about two 

 millimetres pressure, which is the most convenient pressure to work at. 



The vacuum tube, shown in PI. XVI., is made entirely of quartz, and of 

 the same form as the glass tubes used by Pliicker and Hittorf, but with open 

 ends. To one end is attached a glass stopper, through which one of the 

 platinum electrodes passes, and to the other is attached a glass tube and stop- 

 cock, through which the second electrode is introduced and the tube is 

 exhausted, the junction between the glass and quartz being effected simply 

 with an ordinary piece of rubber tube wired on. The connection to the 



